In the storage model of current Network Attached Storage (NAS), a file system is built on several hard disks or several specific storage devices. In this case, the management of the storage device is relatively simple.
Generally, the storage device is managed by obtaining an identifier thereof. In the related art, the following methods are proposed for obtaining the identifier of the storage device.
1. For the storage device developed and researched in-house, some Erasable Programmable Logic Device (EPLD) registers are provided to store a version number and a material order number of a Printed Circuit Board (PCB) of the storage device. This kind of storage devices is uniquely indentified by the version number and the material order number of the PCB. However, the above manner is limited to the storage device developed and researched in-house, and the storage device from external procurement or produced by other manufacturers may not be provided with such information.
2. Corresponding to an Electro Magnetic Compatibility (EMC) magnetic-matrix device, a Universally Unique Identifier (UUID) of this kind of storage devices can be obtained through an Application Programming Interface (API). However, this manner is limited to the magnetic-matrix device or some brands of magnetic-matrix devices.
3. Other Just a Bunch Of Disks (JBOD) devices may have Intelligent Platform Management Interface (IPMI) to provide a product serial number. However, many devices do not have IPMI.
A file system in a distributed system of a cloud computing platform is built on heterogeneous storage architecture. Advantages of such design are that the file system may be adapted to various storage devices. Management of the heterogeneous storage device may allow a user to know operation of each storage device on each storage node, such as a device, a logical volume, a Redundant Array of Independent Disk (RAID), a disk and the like. Additionally, the upper layer may select different storage strategies according to the performance difference of respective storage devices. Obviously, the storage device with better function has more opportunity to be selected.
There is a problem with the current heterogeneous storage architecture, i.e., it is difficult to perform unified management of the storage devices. Since the heterogeneous storage architecture includes a plurality of kinds of storage devices, identifiers of respective storage devices are not unified, and it is impossible to obtain identifiers of storage devices through a unified method, so as to manage application programs and databases in a unified manner. Once the storage devices are powered down, fail or are migrated, the management of the storage devices will become more confusing.